[foundation-board] Iconomical Situation

[Rufus Pollock]

Board members have asked me for a report :) The following is quite
long (and not completely proofed as written quite quickly), for which
I apologize. However, I need to give quite a lot of background to
explain where we are now.

Rufus

### Background

We did the initial on WDMMG prototype for Cabinet Office with
Iconomical back in Autumn 2009.

We included them as main partners in bid for 4IP. 4IP were very
resistant about Iconomical as primary design team due do doubts about
their design skills (very justified in retrospect IMO). At the time I
fought hard to defend Iconomical's inclusion because:

a) We'd worked previously with them on a successful prototype
b) We'd already included them as partners in the bid and to exclude
them I thought would be unfair (they'd already committed some time to
developing ideas for next stage)
c) Timeline we were getting from 4IP made it hard to find any feasible
replacement (4IP didn't sign until end of March and wanted us to
deliver first working version by early May. I did, for example talk
with Eric Rodenbeck of Stamen at 4IP's prompting, and he seemed
interested but they weren't available until June)

However, we did add David McCandless to the team (a very good thing!)
and Steve Cummaford (a UX expert). While Liz Turner wasn't initially
entirely happy about Dave McCandless (he was another designer and was
designated as design lead), by the time we actually went ahead she
seemed quite positive seeing this as opportunity to do something good.

In retrospect it was a mistake to have kept Iconomical in. Problems
started quite early.

1. DM (David McCandless) was in Amsterdam in April for another project
and I therefore tried to schedule as much time with Liz and Iconomical
as possible so they could develop a new design spec for phase II
post-election (at the same time there was work on the backend and
polishing up the prototype for completion of phase I due by the
election (these phases relate to those in the contract with 4IP btw).

I got the impression things weren't going great with Liz and DM. But
April was difficult as it was hard to get hold of everyone (DM was
fitting us in around another job at this point -- something we, and
4IP, knew when we signed him). When I did speak with DM on his return
to UK at start of May it was clear he was finding Liz very difficult.
He described getting a 1h lecture from her at each meeting about how
he didn't understand the data, and unsurprisingly found this rather
de-motivating.

2. I had already scheduled a design team meeting for 7th of May in
London with myself, DM, Liz and Steve (UX). I hoped this meeting,
where there would be several of us, might allow us to move forward the
design stuff construtively. The meeting was actually quite good, with
everyone contributing and us ending up with a basic set of designs
that were documented on the wiki. We also managed to deliver phase I
largely on time.

3. DM now duly delivered some mockups from 7th May meeting (though not
all of them -- more below). Work restarted with Iconomical near end of
May to start working on taking static mockups into a dynamic app. We
had regular monday meetings to plan work and first couple seemed to go
quite well. However on 10/06 Liz wrote a long email about "serious
risks concerning WDMMG", occassioned I believe by a debate at that
week's call regarding how to do the bubble chart (they had an existing
design and DM had a new one, and essentially they didn't want to
implement DM's new design ...). This made specific claims:

1. Absence of long term planning, specifications or requirements.
2. Locked out of decision making
3. Responsibility without autonomy
4. Lack of any binding agreement

The first was I think pretty much without merit (I gave a detailed
response I'm not including here) and this was also true of 2 -- they
had plenty of input they just didn't have the final say, and we had
committed as a team to proceed on the basis of (though not wholly
determined by) DM's designs. 3 was like 2. 4 was an important issue
and related primarily, at least from our end, to a lack of resolution
to open-source question.

4. At this point Jo Walsh participated in discussion with them.
Progress was very slow due to lags from Iconomical's end (because, I
now know, they had taken on other work) and more than a month passed
until much progress was made. The idea was a) to resolve open-sourcing
and payment for work done b) Iconomical to deliver a spec for next
stage which would then be signed off.

Background on opensource: From before start of 4IP I had raised
requirement of open-source with Iconomical. However, under pressure to
begin from 4IP, we did commence work without any formal (written)
agreement though I did believe I had extracted a good-faith promise
from them that this would be resolved once they'd had a proper chance
to look at open-source licenses and decide which one they wanted, how
it fit with their business model etc. Again, in retrospect this was a
mistake and not something I would have permitted w/o the pressure from
4IP. This meant open-source issue was ongoing. Finally in July
discussion open-source issue was resolved with them agreeing to
open-source existing work and work going forward.

5. Unfortunately the spec reflected the general tendency of Iconomical
(led here by Liz I believe) to discard most of DM's work in favour of
her own ideas (as an example: the current 'ptolemaic spheres' bubbles
version was resisted by iconomical for 3 months as being 'bad' or not
fully specced while most people who compared it to the original
iconomical version preferred the spheres).

At this point I really regret not stopping work with Iconomical. The
problem was we were now under time pressure from Iconomical and it was
almost August (a very difficult time to recruit new people). In
addition any newcomer would have to grapple with Iconomical's code
which I had been assured by them was in no fit state for outsiders
(this was part of the reason they didn't want to opensource ...).

All of this was not aided by the intermittency of DM's presence
(finalize designs for all screens still had not been received) which
was due to being very busy and, I think, the pain of working with
Iconomical (even though there was no longer much direct contact the
whole feeling of constructive collaborative development was clearly
missing!).

6. The experience with Iconomical was often rather bi-polar: in that
some phone meetings (esp those with Dave Boyce around!) everything
would seem to be resolved/positive but the next thing would be a
threat to quit or a completely new set of designs with little relation
to what we had done already. Also, as had been case previously they
constantly missed deadlines to deliver specs etc. This had negative
consequences for the project as a whole and for their involvement.

7. Long and short of it was we continued to work with him with aim of
closing everything out for phase II by mid Sept. From this point on
there were less real disputes but it was clear Iconomical did most
stuff reluctantly and without any great interest. I met with them in
person late Sept in Amsterdam and again had a rather bi-polar
experience (promising one moment then back at square one the next).
They continued to resist designs generated by DM (and generated
jointly back on 7th May). Finally we did get a fairly complete set of
designs delivered by early Oct.

8. The problem though was that designs really need polishing. At this
point Iconomical seemed to take zero pride in their work (one assumes
on the logic that it wasn't "their" design) and hence there was
endless need for long emails detailing niggling issues that they
hadn't checked out. There were also significant lags in response which
made things difficult.

### Current situation

At this point there still remain minor issues with the visualization
and having not access to the source code I don't even know whether the
code builds, has a basic README etc. If we pay these things will
*clearly* not get fixed. Furthermore, our agreement in July had been
for open-sourcing going forward not a dependence of open-sourcing on
payment. I have repeatedly asked since August for them to push code
more regularly and was told at the time by Dave that he was just too
busy to get round to it (not that it was a depedency on payment).

In addition the delay in open-sourcing has already been a problem
(several community members wanted to work with the flash but
couldn't). A large contributor to this delay is their lack of
cooperativeness and general delay in responding. They still haven't
replied to my email of the 29th of Nov while I respond within a day.
I've also suggested talking via skype to take things forward without
success.`

Moreover there is a clear lack of support from them for project going
forward so it is probably inevitable that we will hae to rewrite from
scratch. Of course ongoing participation is not something that was
contractually promised but it means we will incur significant expense
to replace their code because of delay and lack of cooperativeness
from their end.

While I believe we *should* pay Iconomical for their work there are 2
fundamental questions:

1. Do we pay before code is fixed and checked (re usuability by third
parties -- I'm afraid of an dump it over the wall setup here)
2. Should there be deductions for the costs in delay (need to rewrite
going forward) caused by Iconomical's behaviour